Eric Bogatin's Blog: What I Learned This Month

Next No Myths Allowed Webinar, “Selecting Capacitor Values for the PDN”, Dec 9  2009, 1 pm EDT. Signup now.

This question is actually one of the topics covered in our upcoming webinar, mentioned above. Of course, the bulk capacitors are used to provide low impedance at a frequency where the VRM is no longer able to keep the impedance of the PDN below the target impedance.

How much bulk capacitance is required depends on the target impedance and this frequency. When the VRM impedance exceeds the target impedance, the impedance of the bulk capacitor (shown in red) had better be there to keep the PDN below the target impedance. This figure is from our Essential Principles of SI (EPSI) class.

Of the possible options, the correct answer is the second one, “the frequency where the VRM impedance is too high.”

This was a pretty easy question as most folks got it correct. It is interesting that 25% selected the third answer, that the bulk capacitance depends on the loop inductance of the bulk capacitors.

The loop inductance of the bulk capacitors is important, as it will influence the parallel resonance with the ceramic capacitors, but it by itself will not determine how much bulk capacitance is required.

This whole topic of selecting capacitors- the number and their value- is very confusing, depends on a large number of system features and details and is difficult to generalize from what worked on one design to the next design.

That’s why we decided to cover this topic in the next webinar. I’ve collaborated with Larry Smith on this project, since he is one of my PDN gurus and I’ve learned a lot about PDN design from him. I hope you will join us on Wed, Dec 9, 2009. If you miss this live event, the webinar will be recorded and posted on our web site.

If you would like to get additional details on PDN design, you can also read the new chapter I wrote in my book, Signal and Power Integrity- Simplified.

And, if you are looking for a new challenge, check out the latest pop quiz we added to the web site!

I hope to see you in cyberspace!

Next No Myths Allowed Webinar, “Selecting Capacitor Values for the PDN”, Dec 9  2009, 1 pm EDT. Sign up now.

The last pop quiz posted on our web site, was a real test of how well you understand differential impedance and odd mode impedance. If you watched our last free webinar, the answer to this pop quiz would have been obvious.

By definition, the odd mode impedance of a line that is part of a differential pair is always ½ x the differential impedance. If the differential impedance is 100 Ohms, the odd mode impedance of either line is 50 Ohms.

In this question, the single-ended impedance is designed as 50 Ohms. How do you have a single-ended impedance and odd mode impedance of 50 Ohms?  It can only be if there is no coupling between the two lines that make up the pair.

As the coupling increases, the single-ended impedance won’t change, but the odd mode impedance will decrease. The only correct answer to this question is, uncoupled.

It is interesting that only 42% of you got it correct! This means 58% got it wrong. This way of specifying differential impedance: a single-ended impedance and a differential impedance, is a common way of specifying line impedance. Now you know, it’s really specifying uncoupled lines, with a spacing about 3x the line width.

To really understand differential impedance, check out the last webinar I did on “Stack up Design for Differential Pairs”. It’s free!.

Next public classes: Essential Principles of Signal Integrity, Advanced Signal Integrity Design, Multi Giga Bit Design: Sept 29- Oct 7, 2009 in San Jose, CA

Next No Myths Allowed Webinar, “Stack-up Design for Differential Pairs”, presented free on Sept  16, 1 pm EDT.

Last month’s pop quiz question related to the impact from coupling on the differential impedance of symmetric stripline differential pairs.

If two coupled symmetric stripline traces start out with a differential impedance of 100 Ohms, and they are brought closer together, what happens to the differential impedance?

The general answer is that increasing the coupling will always decrease the differential impedance; but by how much? The only way to really know is with a 2D field solver. Because I’ve looked at similar problems a lot, I know the answer is somewhere between a 10% and 20% reduction in the differential impedance, but I can’t pin it down any finer without putting in the numbers with a field solver.

My favorite and simplest to use 2D field solver is Si9000 from Polar Instruments. I set up the problem as shown in the figure to the left and swept the spacing to look at the impact on the differential impedance.

In this example, the line width is 5 mils, half ounce copper and the dielectric thickness above and below is 6.7 mils, with a Dk of 4.2. When the spacing is 30 mils, 6x the line width, the differential impedance is 100 Ohms.

As we decrease the spacing, the differential impedance decreases, as we expect. When the spacing is 5 mils, equal to the line width, the differential impedance has dropped to 86 Ohms. This is a 14% reduction in the differential impedance.

We had 325 answers to this quiz on our web site. More than half of you got the right answer, which was either 10% or 20%.

If you want more details about this topic, be sure to check out the next No Myths Allowed Webinar. “Stack up design for differential pairs.” coming up on Sept 16, 2009, or our live class, Essential Principles of Signal Integrity.

Be sure to visit our home page to answer the new pop quiz!

See you in cyberspace.

Check out our next public classes: Essential Principles of Signal Integrity and Advanced Signal Integrity Design, Oct 11-14 in Hillsboro, OR.

Check out our next No Myths Allowed Webinar, “Link Analysis with Return Path Discontinuities”, presented free on July 7, 1 pm EDT.

The pop quiz this month was, “When two adjacent signal lines transition from one signal layer, through a pair of planes, to another signal layer, the return current flows between the cavity formed by the planes. The impact of the return path discontinuity is strongest on which S-parameter term.”

At 28%, the consensus was S11, followed by SDD11 at 23%. This is surprising, as neither answer is correct. This topic is covered in detail in the SI-Insight for June, 2009, released this month, “Ground bounce in Vias.” We also touch on this topic in this month’s webinar, NMA-820, “Link Analysis with Return Path Discontinuities,” which, if you missed the free live event on July 7, can be viewed from the recording on the web site.

Which quality of the pair of lines is most affected by the return path discontinuity? The impedance discontinuity, which S11 is most sensitive to, is minor. And, as we show in the paper and the webinar, the differential signal is mostly immune to this return path discontinuity.

In fact, the biggest impact is on the single ended cross talk between the two lines. The correct answer to this month’s pop quiz is S31, the near end cross talk. Especially when the adjacent lines are far apart, the edge coupled cross talk can be very small, but the cross talk between the signal vias that pass through the cavity can be very large and long range.

The figure to the left, Figure 19 taken from the SI-Insights report, shows the measured near end cross talk between two signal lines, far apart for three design cases. The red measured response is with no signal vias, just two 50 Ohm lines about 15 lines widths apart. The green response is the measured near end crosstalk between these two lines routed from the top layer to the bottom layer, going through vias, with adjacent return vias.

The blue trace is the measured response of similar via transitions, but without the return vias. The coupling between the signal lines from noise injected in the cavity is more than 20 dB higher than with either no vias or vias with adjacent return vias.

Feel up for the challenge of a new pop quiz? Check out the new pop quiz on the web site.

Our next No Myths Allowed webinar, July 7, 1 pm EDT, “Link Analysis with Return Path Discontinuities”. Details and registration available at www.beTheSignal.com

The last pop quiz was: True or false: getting the differential S-parameters are difficult because they require a special differential network analyzer.

The correct answer is False. 99.99% of all VNAs are single ended VNAs and they are perfectly fine at measuring differential S-parameters. With a little matrix math, the single-ended version can be converted into the differential version. This is valid for all linear, passive interconnects.

While many of you got the correct answer, a significant faction, 20%, answered True. This is why the webinar we presented on May 6, NMA-810 S-parameters, Signal Integrity and You, was so important. We looked at what are S-parameters, what they tell us about interconnects and how we transform single-ended S-parameters into differential S-parameters.

A 4-port VNA is really the best tool to get the differential S-parameters. We can take the 16 single-ended elements and transform them into the 16 differential S-parameters, which include not just the behavior of the channel to differential signals but also to common signals and mode conversion by the interconnect.

If all you have is a 2-port VNA, it’s possible to measure the differential S-parameters by performing multiple measurements with different connections and with the other ports terminated. Another option is using a balun. This is great if all you want are the differential elements in the upper left quadrant: SDD11, SDD22 and SDD21. The use of a 10 GHz balun from Picosecond Pulse Labs is also described in the webinar.

If you missed the live event, you can catch it in the archives.

If you enjoyed this quiz, check out the new pop quiz at www.beTheSignal.com

Our next No Myths Allowed Webinar: “S-Parameters, Signal Integrity and You”, May 6, 2009 1 pm EDT. Details are at www.beTheSignal.com

Last month’s pop quiz was:

“To build a transparent differential via, what is the most important feature to engineer?”

Here are the results from 180 participants.  While the most common answer to all signal integrity questions is “it depends”, it’s not always the best answer. In the case of transparent vias, the limitation is really the via stub. The correct answer is the fourth one, minimize the stub length, which 23% of you correctly answered.

All of the other factors are important. but the one with the biggest impact and which limits the bit rate of signals transmitted down the interconnect, is the length of the via stub. As a rough rule of thumb, the maximum stub length, in mils, that can be used in an interconnect system to transmit a bit rate, BR, in Gbps, is roughly:

Len < 300 mils/BR.

For the details on this and other properties of differential pairs, check out the last No Myths Allowed webinar, “NMA-800 Practical Differential Pair Design”. The handouts are available for download, and if you missed the live webinar, you can view the recording from our web site, www.beTheSignal.com.

And don’t miss our next No Myths Allowed webinar, “S-Parameters, Signal Integrity and You” on May 6 at 1 pm PDT.  I’ll see you there!

Last month’s quiz was a very fundamental question in signal integrity. “The open circuit output voltage of a driver is 1v. Its output impedance is 5 ohms. It is driving a 50 ohm transmission line. What is the initial voltage launched into the transmission line?”

This question is really about how to think of a transmission line. Once you realize that when the signal looks into the transmission line, the initial impedance it sees is the characteristic impedance of the line, then the answer will be obvious.

Initially, the transmission line looks like a resistor of 50 ohms. With this perspective, the signal sees a voltage divider, with 5 ohms in series with the 50 ohms of the transmission line. The voltage launched into the line is just 1 v x 50 ohms/55 ohms = 0.91 v.  The answer was number 2.

As you can see on the results table to the left, most of you got the right answer. For those that want to understand this better, take a look at OLL-115 and BTS002.

In addition, we’ve just launched our first hands on labs. The first one, HOL-101 allows you to explore this question of the voltage launched into a transmission line. I invite you to take it for a spin. We love feedback, so let me know what you think of this type of hands on lab. I have plans for many more.

Check out the latest quiz which is about our new webinar series. See you in cyberspace!

This pop quiz question on www.beTheSignal.com, was really about estimating the equivalent series inductance (ESL) of a decoupling capacitor and how low it can be engineered. Of course, the ESL is all about how the capacitor is integrated into the board and the stack up of the board.

As a rough approximation, the ESL of a capacitor is composed of three elements, the sheet inductance of the capacitor and surface traces, the via loop inductance to the top of the power and ground plane cavity and the spreading inductance in the power and ground plane cavity. Each of these can be estimated with simple approximations.

The sheet inductance of the surface traces and the capacitor body is 32 pH/mil x h1 x Len/width

where

h1 = distance from top surface to top of the power-gnd plane cavity, in mils
Len = the total length of surface traces and capacitor body, in mils
width = width of the surface traces, in mils

The ratio of Len/width is really the number of squares of surface trace, and 32 pH/mil x h1 is the sheet inductance per square between the board surface and the top of the power-gnd cavity. If h1 = 5 mils, the sheet inductance of the top layer is 160 pH/square. If there are 5 squares of surface traces, this is a total loop inductance from the surface traces of about 800 pH.

The loop inductance of the vias to the top cavity is about 15 pH/mil. For a total of 5 mils, this is about 75 pH.

Finally, the spreading inductance in the planes is about 20 pH/mil x h2 x ln(B/D), where
h2 = the distance between the power and ground planes, in mils
B = the distance between the capacitor and the BGA package pins, in mils
D = via diameter in mils.

If h2 = 3 mils, a common minimum thickness in all fab houses, and B = 500 mils and D = 13 mils, the spreading inductance is roughly 20 x 3 x 3.6 = 220 pH

Without doing anything special, the typical ESL of an 0603 multi layer ceramic capacitor (MLCC) might be on the order of 800 pH + 75 pH + 220 pH = 1.1 nH. The choices were 0.2 nH, 1.5 nH and 5 nH. the closest to our simple estimate is 1.5 nH.

Of course, if you do everything right, bring the power-gnd plane cavity close to the surface, use a minimum number of squares for surface traces, use a thinner power-gnd plane cavity, the ESL can be dropped even more. If you use interdigitated capacitors, like the X2Y, the parallel combination of capacitor elements can reduce the surface trace inductance and reduce the spreading inductance in the cavity to achieve an ESL on the order of 0.2 nH.

Check www.beTheSignal for the latest online lectures on inductance and for the next pop quiz!

Pencils down! The results of the latest pop quiz are in. The question posted on www.beTheSignal.com a few weeks ago was “A data stream has a 10-90 rise time of about 300 psec. What is the approximate bandwidth of the signal?”

There were 303 answers submitted. Our web site assures that a person can submit only 1 answer.  70% of you got it correct.

The bandwidth of a signal is the highest sine wave frequency component in eh signal. As a pretty good approximation, the bandwidth of a signal is about 0.35/RT, where RT is the 10-90 rise time.

In this example, the rise time is 0.3 nsec, so the bandwidth is about 0.35/0.3 = 1 GHz. Of course, this assumes a linear or gaussian edge shape. If it is grossly distorted, or has a long tail, such as with a lossy line, the 10-90 rise time can be long, and this approximation would be an under estimate.

The use of the term bandwidth is always only a rough approximation. If it is important whether the bandwidth is 1.1 GHz or 1.3 GHz, don’t use the bandwidth, use the entire spectrum. Want to know more about rise time and bandwidth, see chapter 2 in my book, or attend the EPSI class.

For you puzzle geeks, a new pop quiz has been posted! The answer will appear in this blog in a few weeks.